Acceptance rates: what your payment declines really say about your stack

A stagnant acceptance rate isn't the bank's fault. It is almost always a symptom of a payment stack that has run out of steam, built around a single PSP with no plan B for when the acquirer falters or the issuer tightens its rules.

The problem is that these declined payments don't show up in your P&L. There is no "lost revenue" line item on the income statement. Yet, every failed transaction represents a validated cart, a convinced customer, and marketing budget already spent, with zero revenue to show for it.

The scale of the impact is eye-opening. For a site processing €10M in annual volume, gaining 1 percentage point in authorization rate brings in approximately €100,000 in net revenue. Without a single line of code on the acquisition side, without an extra euro in SEA, and without touching your checkout conversion rate.

In other words, before chasing additional traffic, there are often ten to fifteen points of acceptance to be recovered in the plumbing. That is exactly what a layer ofpayment orchestrationenables, and that is the subject of this article.

Auditing your flows: the 4 KPIs you shouldn't confuse

Most teams talk about "acceptance rate" as if it were a single figure. In reality, there are four, and they don't measure the same thing. Confusing them means flying blind.

A small detail that many overlook: you must deduplicate attemptsA customer who tries three times before succeeding counts as one sale but three requests. Without a central platform to clean up these duplicates, your apparent decline rate is artificially inflated, and your decision-making is based on flawed data.

The payment expert's take. On a monthly volume of €500,000 with an 80% acceptance rate, you are leaving €100,000 on the table. Boosting your stack to 95% means an extra €75,000 coming in the following month. Payments are not a cost center to be squeezed; they are an acquisition channel to be activated.

Hard declines vs. soft declines: the basics many miss

Not all declines are created equal. The ability to sort them in real time is what separates modern stacks from aging integrations.

Hard declines: don't push your luck

These are definitive refusals. Code 41 and 43 for stolen or lost cards, code 54 for expired cards, code 14 for invalid numbers (you can find a list of these codes documented by Stripe). In these cases, retrying the transaction is counterproductive. Visa and Mastercard monitor your approval ratio. Too many retries on dead cards, and you end up in monitoring programs that will weigh down all your future payments.

The rule is simple. Your stack must identify these codes, block the retry, and immediately offer the buyer another payment method—such as a digital wallet, instant bank transfer, or BNPL, depending on the context.

Soft declines: 80 to 90% of the opportunity

Insufficient funds at the start of the month, weekly limits reached, temporary issuer server downtime, or false-positive fraud flags. These declines are recoverable, provided you have a smart retry logic in place.

The right approach is rarely to retry within the minute. Retrying 24 hours later, via a different acquirer or with a different authentication level, works significantly better than retrying two hours later. Well-calibrated setups recover up to 20% of declined transactions, without any human intervention.

False positives: the silent hemorrhage

The worst of the lot. A solvent buyer, ready to pay, rejected because of a silly payload anomaly. An extra space in the name, a date in US format instead of European, a poorly mapped BIN, and it's an immediate rejection.

According to Checkout , the revenue loss attributed to false declines has surged by 300% since 2019 in Western e-commerce. And the impact doesn't stop at the abandoned cart. Nearly one in two major accounts finds that these false declines permanently damage retention. A buyer wrongly rejected rarely tries a second time on the same site.

TCO and technical debt: the single PSP trap

Relying entirely on a single PSP for your acceptance means creating a SPOF (single point of failure). The day the partner's API goes down, you stop billing. Not in an hour. Instantly.

Beyond the operational risk, there is the question of total cost of ownership. Building and maintaining your own orchestration layer in-house remains a heavy project that few CIOs have the resources to absorb.

Delegating complexity to an orchestration layer frees up tech teams for what creates value on the product side, and guarantees ROI within the first few months.

Smart routing: routing every transaction to the best acquirer

Orchestration decouples the e-commerce layer from financial providers. For every transaction, the smart routing engine reads the context in real time (card BIN, buyer's country, currency, average basket, acquirer's historical success rate for this profile) and sends the request to the highest-performing processor at that exact moment.

If the primary acquirer returns a technical error or goes down, the fallback automatically reroutes the transaction to a backup acquirer, completely transparently to the buyer.

Auchan Retail case study. By centralizing its payment acceptance and activating multiple alternative payment methods through an orchestration layer, the retailer gained 2 percentage points in acceptance rate on €1 billion in web transaction volume. That is €20 million in additional revenue generated directly from the technical stack, at constant volume.

3DS exemptions and network tokens: the two game-changing levers

Two components now make the difference between a modern stack and an obsolete one.

Authentication exemption (TRA). Under PSD2, the rules engine evaluates the risk of each transaction in milliseconds. For trusted profiles (repeat buyers, consistent basket sizes, domestic BINs), it requests an exemption from the issuer for 3DS. The buyer proceeds without friction, security remains intact from a risk perspective, and the conversion rate increases automatically.

**Network tokens.** Visa and Mastercard replace the PAN with an encrypted token, unique to each merchant. This has three cumulative effects. An increase in the authorization rate of up to 4% according to VisaA clear reduction in fraud, because the token is for your use only. And above all, the automatic updating of expired or reissued cards, which eliminates involuntary customer churn on subscriptions and customer accounts.

Local payment methods: the key to international growth

Selling in Europe without iDEAL in the Netherlands, Bancontact in Belgium, or Twint in Switzerland means missing out on a huge share of the market. Buyers won't adapt. They will simply go to a competitor that offers their preferred payment method.

An API-first architecture allows you to activate a new local method from a back-office in just a few clicks, without a three-month IT project.

Vertbaudet case study. The brand replaced a dozen PSP integrations with a single API connected to 8 acquirers. Results in local markets: Twint at 50% market share in Switzerland, Bancontact at 60% in Belgium, +4 points in credit card acceptance and +10 points in frictionless checkout flows.

Reconciliation and observability: bringing finance back into the loop

Multiplying PSPs also means multiplying reporting formats, payout schedules, and accounting entries. If orchestration stops at routing, the finance department is left to piece everything together by hand.

A good platform consolidates transactional data into a single vault and exports entries in the formats required by ERPs (CSV, MT940, CODA). By giving the finance department the right tools to ensure financial data reliability and automate reconciliation, the turnaround time drops from days to hours, and observability becomes real-time.

iQera case study. Centralizing multiple PSPs has simplified collections and streamlined the group's accounting. Adding PayPal and then pay-by-link was done without any IT project. The result: +2 points in net acceptance and measurably more effective debt recovery.

PSP or agnostic orchestrator: the real strategic question

This is the trade-off that will weigh on the next five years. Staying with a traditional PSP means accepting vendor lock-in : their SDK, their checkout flow, their pricing terms, and the difficulty of leaving once your volumes become significant.

An agnostic orchestrator does the opposite. It abstracts technical complexity without imposing any proprietary acquisition flow. You retain control over your acquiring contracts, your data, and above all, the lever that really matters in the long run: competitive bidding. Having three acquirers competing for your volume means gaining several dozen basis points on every transaction, and a stack that gets stronger with every new addition instead of becoming more rigid.